2534
Vol. 48
C. H. PRESCOTT, JR.
t o prevent actual diffusion during the run. Having made the run with iodine a t room temperature, the measurements were carried up to the higher temperatures, preferably exactly that of the air run. Preliminary results are: T (absolute) = 300"; Pair= 0.242 mm.; Pair I2 = 0.448 mm.; T = 908";Pair I2 I = 0.825 mm.; T = 925"; Pair IZ I = 0.847 mm.; T = 1213";Pair I = l.5SO mm. I n these runs the temperature of the furnace was determined by a calibration using the instrument as a gas thermometer. This calibration checked that using a thermocouple to within a few degrees except a t the higher temperatures where there was reason to suspect that the thermocouple (which was a very imperfect one) was a t fault. We should have liked t o check this calibration optically, but the necessary pyrometer was not available a t the moment. The most serious source of error came from the fact that the water used t o cool the furnace was a t a temperature lower than room temperature and therefore it was suspected that iodine was being deposited a t the point (A). There was, therefore, considerable chance of the partial pressure of the iodine in the cold gas changing with the temperature of the furnace. Indeed, a t the higher temperatures the'pressure change is higher than that calculated on the assumption of complete dissociation of the iodine. However, the results a t about 915" lead to a value for log k of 3.7 (Lewis and Randall calculate 3.4 * 0.2).
+
+ + +
+ +
Summary A new method of measuring the temperature and molecular weight of gases has been described. A preliminary determination of the dissociation constant of iodine a t 740" has been made. BALTIMORE, MARYLAND [CONTRIBUTION FROM THE
GATESCHEMICAL
LABORATORY, CALIFORNIA INSTITUTE OF
TECHNOLOGY, No.
1061
THE EQUILIBRIUM BETWEEN ZIRCONIUM OXIDE AND CARBON AND THEIR REACTION PRODUCTS AT INCANDESCENT TEMPERATURES BY C. H. RECEIVED JUNE 28; 1926
PRESCOTT, JR.
PUBLISHED OCTOBER6 , 1926
Plan of the Investigation This research was begun in the fall of 1923 a t the suggestion of Professor A. A. Noyes as an attempt to devise an apparatus and precision technique for the study of chemical equilibria a t incandescent temperatures. The particular subject of interest was the investigation of the stability of refractory oxides. The reduction of zirconium oxide by carbon was chosen as the immediate subject for research.
Oct., 1926
ZIRCONIUM OXIDE AND CARBON
2535
This reduction has been studied by Troost, Moissan,2and more recently by Wedekir~d.~All three report a carbide as the product, Moissan and Wedekind agreeing on the formula ZrC. Van Arkel and deBoer4 prepared the carbide by passing zirconium tetrachloride, hydrogen and carbon monoxide over a hot tungsten filament which precludes the presence of free carbon, and checked their results with x-ray powder photographs. They report that the formula is ZrC and that the substance has a sodium chloride type of structure. The formula ZrCz reported in the International Critical Tables seems to be without justification. The predominant feature of the present technique is the small size of the furnace. This allows the attainment and very accurate control of high temperatures with a minimum of power. It also makes it possible to enclose the apparatus completely in glass, and the energy evolved is not so great as to prevent thermostatting the furnace and so obtaining uniform conditions in the environment. The furnace is comparatively easy to manufacture and replace, which is essential as it is bound to deteriorate a t the temperatures employed. The limiting factor in the present technique is the temperature control, by pyrometer measurements, in the face of a fluctuating line voltage. The use of commercial power is not to be recommended. The chief fundamental limitation of the method is that it applies directly only to equilibria with one gaseous component, where the reaction is not limited by gaseous diffusion but progresses by direct displacement. Acknowledgment is due t o the Norman Bridge Laboratory of Physics for the loan of the optical pyrometer, t o Dr. W. E. Forsythe and Dr. A. G. Worthing of the Xela Research Laboratory for assistance with the pyrometry, and to Dr. S. B. Hendricks for the x-ray work on the solid phases. This investigation was aided financially from a grant made to Professor A. A. Noyes by the Carnegie Institution of Washington. The Temperature Measurements The temperature was measured with a Leeds and Northrup aptical pyrometer fitted with an auxiliary lens of 4 diopters which enabled it to be brought to a distance of 20 cm. from the furnace. At the close of the experiments this instrument was recalibrated against a standard lamp ohtained from the Nela Research Laboratory. At the same time the transmission was measured of a water cell made of a section of a 1.5-liter Pyrex flask and a piece of window glass, which closely approximated the conditions of the furnace and thermostat windows with water between. Troost, Compt. rend., 116, 1227 (1893). Moissan and Lengfeld, ibid., 122, 651 (1896). Wedekind, Chem.-Ztg., 31, 654 (1907). Van Arkel, Physka, 4, 286 (1924). Van Arkel and deBoer, Z. anorg. Chem., 148, 345 (1923).
* Moissan, ibid.,116, 1222 (1893).
2536
C. H. PRESCOTT, JR.
Vol. 48
The corresponding correction t o the observed temperature was calculated from the Wien radiation law by the formula5 (1/S) - (l/O) = (A log t ) / 0.4343~2,where O is the observed temperature, S is the brightness temperature of carbon, t is the transmission of the water cell (83.9%)' X is the effective wave length of the pyrometer color screen (0.66p), and c2 is 14330,~ degree. Since the new calibration agreed within 3' with the old Bureau of Standards calibration in the temperature range of the equilibrium measurements, corrections for the emissivity of carbon were used as determined on the old hole in the center of the wall had been scale. A furnace with an 0.8". mounted in furnace bulb No. 1 and the apparent brightness temperature of the adjacent wall compared with the temperature of the hole a t a series of temperatures from 1250'K. to 2050'K. This correction is T - S = 0.017 S-8 degrees. During the calibration it was found that readings could be duplicated t o about 1.5'. The emissivity correction probably introduces errors not greater than 4", so the temperatures as reported are probably relatively good to 2' or 3', and absolutely to 6' or 7". Apparatus and Procedure The furnace, shown in Fig. 2 (where the dimensions are stated in inches), is a tube of Acheson graphite 5.7 cm. long, 0.635 cm. in diameter a t the ends, and 0.318 cm. in the central section for a length of 2.54 cm., with a 0.04cm. wall. This is supported on tungsten rods pressed into the ends of the tube. The charge is retained a t the center by graphite plugs resting loosely against the tungsten rods. The charge was in every case a pellet of compressed zirconium oxide and carbon, approximately 0.64 cm. long and 90 mg. in weight, fitting Fig. 1.-Furnace bulb No. 1. loosely into the furnace. The zirconium oxide was made by igniting Kahlbaum's zirconium nitrate. The carbon was Acheson graphite, the same as that of the furnace. Furnace bulb No. 1 (Fig. 1) was found inadequate for equilibrium measurements, but was used to determine the emissivity of carbon and to run samples to completion in a vacuum. The furnace was horizontal, the tungsten rods resting in steel guides clamped to water-cooled copper leads. Electrical connection was made by flexible copper wire fastened t o steel 6 W. E. Forsythe, J . Optical SOC. Am., 4, 305 (1920); 5 , 494 (1921). Report of Standards Committee on Pyrometry.
Oct., 1926
ZIRCONIUM OXIDE AND
CARBON
2537
clamps on the ends of the tungsten rods. The outer ends of the watercooled leads were tapered to fit ground-glass seals in a Pyrex plug which in turn was ground to fit a l-liter Pyrex flask. Seals were made with Dennison's banker's specie sealing wax, and the flask was permanently connected to the vacuum system. I n furnace bulb No. 2 (Fig. rods were welded t o flexible - 2 ) the tungsten copper wire through a nickel intermediary. The copper was melted completely over the nickel to prevent the formation of nickel carbonyl. The flexible cord was then welded to a 0.635cm. copper rod which was silver soldered through a copper disk seal in the end of a 1.9cm. Pyrex tube. The central section was a 2.54cm. tube with a window mounted 3 cm. from connecting the axis, and a 2". tube. The extreme fragility of the furnace necessitated the flexible leads as well as the following assembling procedure. The lower lead was clamped vertically in a special support, and the furnace pushed onto the tungsten rod. Then the central tube was placed over the furnace and down against the seal. The top lead was then clamped somewhat above its final position and lowered gradually as the furnace was worked onto the tungsten rod. Finally, the cenc tral tube was raised into position and connected with a hand torch. Support, furnace and all were Fig. 2.-Furnace bulb No. 2 and furnace. then placed in position and the connecting tube was sealed to the rest of the apparatus. The furnace was connected (see Fig. 3) through a small trap T to the manometer M which had a range of 2 meters, and a piece of capillary tubing at the bottom of the column t o damp out oscillations a t high pressures. The outlet from the Toepler pump P and the connecting tube to the stopcock A comprised the rest of the enclosed gas space during measurements, which was all placed below the water level in a thermostat. This
r-31 --__
2538
Vol. 48
C. H . PRESCOTT, JR.
was of 300-liters capacity and equipped with a 500-watt heater and an electric fan a t opposite ends of the same relay, and two stirrers, one located near the furnace so as to pass by i t a constant stream of water from below. It was set a t 35' and could dissipate an extra kilowatt of power without rise of temperature. The thermostat was set on a laboratory desk, and had a glass window through which the temperature measurements were made.
60Cm.
l'h 7
Fig. 3.-Apparatus
for high-temperature equilibria.
Current was supplied by a l-kilowatt transformer with an 8.5ohm control resistance in the primary. The furnace current was of the order of 150 amperes a t 6 volts. I n Fig. 3 the essential features also of the apparatus outside the thermostat are shown: a Toepler pump P t o force gas into the furnace bulb, a gas reservoir R, and a stopcock B which isolates them from the McLeod gage G and the vacuum pumps which are connected a t V, a single-stage
Oct., 1Q2G
ZIRCONIUM OXIDE AND CARBON
2530
mercury-diffusion pump backed by a Hyvac oil pump. The top of the manometer is also brought back to the high-vacuum line so as to give absolute pressure measurements. Due t o the high pressures sometimes used, the delivery tube and compensating arm of the Toepler pump extend upwards to the same height as the manometer. The stopcocks A, B and C were set in sealing wax and warmed with a hand torch when turned. A and B were of lcm. bore t o facilitate evacuation and prevent clogging. They withstood differences of pressure up to 2.5 atmospheres and remained tight for many turnings. After the furnace was assembled and installed and the thermostat filled, the whole system was evacuated with Cocks A and B open, C closed and Pump P connected to the fore vacuum. When a pressure of mm. was obtained, it was assumed that the furnace was tight, and the furnace was run for a few minutes to drive off any adsorbed inert gases. Then C was opened, and when vacuum was reestablished, B was closed. Carbon monoxide was generated from sodium formate and sulfuric acid, washed in sodium hydroxide and sulfuric acid, passed through phosphorus pentoxide, and admitted to the apparatus through D, a greased three-way stopcock. Simultaneously, the vacuum was released on the pump P. The apparatus was filled, usually to one atmosphere, but several times to one-half atmosphere as the pressure in the reservoir had to be below that in the furnace. Then Cocks C and A were closed. Additional gas was pumped into the furnace bulb by applying pressure to P from the compressed-air line or an oxygen tank. To set the pressure precisely, the Toepler pump was operated with the furnace running. The pressure could be reduced to that of the reservoir by opening the cock A. As the thermostat had t o be drained a couple of inches whenever this cock was turned, runs were made as much as possible in the order of increasing pressure. The rate of reaction was so slow, and the quantity of material so small, corresponding t o a total change of 30 cm. pressure in the enclosed volume of 100 cc., that it was impracticable to wait for the system to attain equilibrium. Instead, it was merely used as an indicator of the distance from equilibrium. Each run constituted a measurement of the rate of reaction a t a predetermined temperature and pressure. The current was first thrown on and the furnace brought to the approximate temperature. Then gas was pumped in to the proper pressure. Readings were next made on the optical pyrometer and the furnace current adjusted till the temperature was that set for the run. Then the pressure was read on the manometer. The temperature was adjusted before each of the pressure readings, which were taken a t intervals of three t o ten minutes for periods of one to three hours. The temperature fluctuated considerably, commonly 5 O , often loo, occasionally as much as 30°, as a result of the rapid changes in line voltage.
2540
C.
Vol. 48
H. PRESCOTT, JR.
With the apparatus and procedure described above it was found impossible to obtain interpretable results until allowance was made for the emission of carbon monoxide produced by the oxygen adsorbed on the carbon furnace. With a blank pellet of powdered graphite, rates of emission were found entirely comparable with previously supposed rates of reaction, increasing with the temperature, but entirely independent of the pressure. After baking for five hours a t 2000' and 1.5 atmospheres]
-
--3b&- -
37.0(--
-
36.01-
--
0.0Y5-Y
0 0 0
I
b0m;s.
yo.r
